Nitrogen Dioxide Inhalation Exposures Induce Cardiac Mitochondrial Reactive Oxygen Species Production, Impair Mitochondrial Function and Promote Coronary Endothelial Dysfunction

Traffic air pollution is a major health problem and is recognized as an important risk factor for cardiovascular (CV) diseases. In a previous experimental study, we showed that diesel exhaust (DE) exposures induced cardiac mitochondrial and CV dysfunctions associated with the gaseous phase. Here, we hypothesized that NO2 exposures to levels close to those found in DE induce a mitochondrial reactive oxygen species (ROS) production, which contribute to an endothelial dysfunction, an early indicator for numerous CV diseases. For this, we studied the effects of NO2 on ROS production and its impacts on the mitochondrial, coronary endothelial and cardiac functions, after acute (one single exposure) and repeated (three h/day, five days/week for three weeks) exposures in Wistar rats. Acute NO2 exposure induced an early but reversible mitochondrial ROS production. This event was isolated since neither mitochondrial function nor endothelial function were impaired, whereas cardiac function assessment showed a reversible left ventricular dysfunction. Conversely, after three weeks of exposure this alteration was accompanied by a cardiac mitochondrial dysfunction highlighted by an alteration of adenosine triphosphate (ATP) synthesis and oxidative phosphorylation and an increase in mitochondrial ROS production. Moreover, repeated NO2 exposures promoted endothelial dysfunction of the coronary arteries, as shown by reduced acetylcholine-induced vasodilatation, which was due, at least partially, to a superoxide-dependent decrease of nitric oxide (NO) bioavailability. This study shows that NO2 exposures impair cardiac mitochondrial function, which, in conjunction with coronary endothelial dysfunction, contributes to cardiac dysfunction. Together, these results clearly identify NO2 as a probable risk factor in ischemic heart diseases.


Introduction
Air pollution is a major health problem and is recognized as an important risk factor for cardiovascular (CV) diseases [1]. Among air pollutants, nitrogen dioxide (NO 2 ) is a reactive gas and a primary pollutant originating from a variety of sources, especially from the combustion of fossil fuel and present in diesel exhaust (DE) [2]. This pollutant is regarded as a marker of motorized road traffic pollution and has been associated with CV adverse health outcomes [3]. However, close correlations between NO 2 and other air pollutants, mainly particulate matter (PM), make it difficult to identify adverse effects due to NO 2 alone [4]. Moreover, there is limited experimental evidence from controlled human exposure and animal toxicology studies for NO 2 , which have focused largely on respiratory parameters. It should also be stressed that NO 2 air monitor networks are sparse or non-existent in many countries [5]. As a result, associations between NO 2 exposures and CV health may be underestimated.
Some experimental investigations have shown the presence of biomarkers for CV effects, including markers for oxidative stress, inflammation, cell adhesion and endothelial dysfunction after NO 2 exposure in rodents [6,7]. In term of vascular function, a controlled human exposure study did not find any impact on brachial artery reactivity after one h exposure to four ppm NO 2 [8], whereas in a previous study, one hour exposure to diluted DE, including both gaseous and particulate phases, induced a vascular dysfunction and impaired endogenous fibrinolysis in men [9]. In the same way, Lucking et al. [10] demonstrated the preventive action of a particle trap on the vascular and prothrombotic effects of DE inhalation, suggesting the involvement of PM in these effects. Although these studies showed that the acute adverse vascular effects of air pollution are mediated by components other than NO 2 , other studies suggested that NO 2 itself contributes to the deleterious effects of DE [11,12], but the CV functional consequences of these exposures remain to be established.
In a previous study, we showed that DE-induced cardiac mitochondrial and CV dysfunctions were associated with the gaseous phase, since the observed effects were similar upstream or downstream of a particle trap [13]. However, the question arises as to whether NO 2 may be involved in these CV effects. Mitochondrial defect is a prominent feature of most CV diseases, such as heart failure, cardiac hypertrophy, ischemic heart disease and atherosclerosis [14]. Dysfunctional mitochondria might generate excessive levels of reactive oxygen species (ROS) and one of the consequences is a decline in endothelial NO bioavailability [15,16]. In the area of air pollution, mitochondria are a potential target because they have not only key roles in cardiac cell functions, but they also produce ROS that may contribute to oxidative stress, a major determinant for air pollutant toxicity [17,18].
Based on these observations, it is tempting to speculate that NO 2 exposures to levels close to those found in DE induce mitochondrial dysfunction, which contributes to the alteration of coronary microvascular reactivity associated with a cardiac dysfunction. To test this hypothesis, we studied the effects of NO 2 inhalation in rats on: 1. cardiac mitochondrial function and ROS production, 2. cardiac function, 3. endothelium-dependent vasodilator responses of coronary arteries and 4. the role that ROS could play in these alterations. These experiments were performed after acute or repeated exposures, with the aim of establishing an exposure-response relationship and finally of showing the contribution of NO 2 in air pollution-mediated cardiovascular effects.

Materials and Methods
Animals and animal care. This study conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996) as well as European legislation and was approved by the ethics committee CENOMEXA n • 54 (authorization number n • 01796.02). Male Wistar rats 9-11 weeks old were purchased from Janvier Labs (Le Genest Saint Isle, France) and maintained in the animal facility at 21 • C on a 12-h light/dark cycle with free access to water and food.
Whole-body inhalation exposure. Animals were randomly divided into six groups and placed in inhalation chambers during the exposures in the animal facility, as previously described [19]. Four groups were used for acute exposures to 5 ppm NO 2 or to clean air for a single 3 h period and the evaluations were performed after a recovery period of 1 h or 24 h in clean air in the animal facility; two groups were used for repeated exposures to clean air or to 5 ppm NO 2 3 h/day, 5 days/week during 3 weeks and the evaluations were performed at 1 day post-exposure after a recovery period in clean air in the animal facility, which is vented with an high-efficiency particulate air and active charcoal filter to remove any pollutants from the outside. Clean air exposure was conducted using dry compressed air (dew point −70 • C) particle-free (<1 particle/cc using a condensation particle counter) with active charcoal treatment. Relative humidity (RH) is set to 50% RH with a steam humidification control system. Downward Mass Flow Controller is protected using cryotrap and adequate filtering. The experimental design is presented in Figure 1. To comply with the established "3R" (Reduce, Refine and Replace) principles, only the minimum number of animals has been used. At least n = 6 rats per group were used for the echocardiographic and biological assessments and at least n = 4 rats per group were used for the coronary vascular reactivity.
The atmosphere of NO 2 in the inhalation chambers was obtained by mixing NO 2 gas (Air Liquid Product, France) with the filtered ambient air in a dilution column upstream of the inhalation chambers, and continuously measured during the exposures by a chemiluminescence analyzer environmental range (AC31M, Environnement SA, Poissy, France). The calibration of the chemiluminescence analyzer was done using a known concentration of NO 2 . The experimental device to generate the NO 2 atmosphere was placed in a room outside of the animal facility in order to avoid transfer of potentially stressful noise.
The NO 2 concentration is in the range of the concentrations measured in our previous studies conducted with diluted DE emitted under dynamic conditions ("New European Driving Cycle" NEDC) and characterized by an average level of 3.3 ppm NO 2 [13,20]. The choice of 5 ppm NO 2 was also based on previous experimental studies demonstrating effects on cardiac biochemical parameters [6].
Labs (Le Genest Saint Isle, France) and maintained in the animal facility at 21 °C on a 12-h light/dark cycle with free access to water and food.
Whole-body inhalation exposure. Animals were randomly divided into six groups and placed in inhalation chambers during the exposures in the animal facility, as previously described [19]. Four groups were used for acute exposures to 5 ppm NO2 or to clean air for a single 3 h period and the evaluations were performed after a recovery period of 1 h or 24 h in clean air in the animal facility; two groups were used for repeated exposures to clean air or to 5 ppm NO2 3 h/day, 5 days/week during 3 weeks and the evaluations were performed at 1 day post-exposure after a recovery period in clean air in the animal facility, which is vented with an high-efficiency particulate air and active charcoal filter to remove any pollutants from the outside. Clean air exposure was conducted using dry compressed air (dew point −70 °C) particle-free (<1 particle/cc using a condensation particle counter) with active charcoal treatment. Relative humidity (RH) is set to 50% RH with a steam humidification control system. Downward Mass Flow Controller is protected using cryotrap and adequate filtering. The experimental design is presented in Figure 1. To comply with the established "3R" (Reduce, Refine and Replace) principles, only the minimum number of animals has been used. At least n = 6 rats per group were used for the echocardiographic and biological assessments and at least n = 4 rats per group were used for the coronary vascular reactivity.
The atmosphere of NO2 in the inhalation chambers was obtained by mixing NO2 gas (Air Liquid Product, France) with the filtered ambient air in a dilution column upstream of the inhalation chambers, and continuously measured during the exposures by a chemiluminescence analyzer environmental range (AC31M, Environnement SA, Poissy, France). The calibration of the chemiluminescence analyzer was done using a known concentration of NO2. The experimental device to generate the NO2 atmosphere was placed in a room outside of the animal facility in order to avoid transfer of potentially stressful noise.
The NO2 concentration is in the range of the concentrations measured in our previous studies conducted with diluted DE emitted under dynamic conditions ("New European Driving Cycle" NEDC) and characterized by an average level of 3.3 ppm NO2 [13,20]. The choice of 5 ppm NO2 was also based on previous experimental studies demonstrating effects on cardiac biochemical parameters [6]. Cardiac function evaluations. Echocardiographic assessments were conducted blind to the animal group and performed in sedated rats (100-135 mg/kg ketamine; 3 mg/kg xylazine) after different recovery periods. Left ventricular (LV) dimensions, and function were assessed with a Vivid 7 ultrasound device (General Electric Healthcare, France), as previously described [21]. Briefly, cardiac Cardiac function evaluations. Echocardiographic assessments were conducted blind to the animal group and performed in sedated rats (100-135 mg/kg ketamine; 3 mg/kg xylazine) after different recovery periods. Left ventricular (LV) dimensions, and function were assessed with a Vivid 7 ultrasound device (General Electric Healthcare, France), as previously described [21]. Briefly, cardiac ventricular dimensions were measured using M-mode tracings recorded from a two-dimensional short-axis view at the level of the papillary muscles. Echocardiography provided measurements of left ventricular (LV) end-diastolic (LVedd) and end-systolic (LVesd) diameters and posterior wall thickness at diastole (PWEDT) and at systole (PWEST). LV systolic function was assessed by the fractional shortening (FS) [(LVedd-LVesd)/LVedd] x 100. Velocity-time integral was measured by pulsed-wave Doppler, and cardiac output (CO) was calculated as CO = aortic velocity-time integral × [(π × LV outflow diameter)2/4]/100 × heart rate.
Vascular function evaluation. Coronary vascular reactivity was evaluated by myograph (Dual Wire Myograph System; Danish Myo Technology), as previously described [22]. In brief, the heart was placed in cold, oxygenated Krebs buffer. A segment of the septal coronary artery, 1 mm long and 100 µm in diameter, was carefully dissected and mounted in a small vessel myograph for isometric tension recording. All measurements were performed after vessel contraction with 10 −5 M serotonin, and pharmacological inhibitors were applied for 30 min before assessing the relaxant responses. The endothelium-dependent relaxations to acetylcholine (10 −9 to 10 −4.5 M) were assessed in the absence and in the presence of the NOS inhibitor Nω-nitro-l-arginine (L-NNA; 10 −4 M) or superoxide dismutase (SOD; 200 UI/mL). Endothelium-independent relaxations to the NO donor sodium nitroprusside (SNP; 10 −9 to 10 −4.5 M) were assessed.
Cardiac mitochondrial evaluations. For the mitochondrial evaluations, two types of sample were freshly prepared from LV as previously described [13,23]: permeabilized cardiac fibers for evaluations of mitochondrial oxidative phosphorylation capacity (OXPHOS) and isolated mitochondria for ATP and ROS measurements. The OXPHOS capacity was evaluated in situ on permeabilized cardiac fibers, in order to maintain the mitochondria in their normal intracellular position and their interactions with other organelles. OXPHOS was measured polarographically at 22 • C using a Clark-type oxygen electrode (Strathkelvin Instruments, Scotland, UK). Briefly, permeabilized cardiac fibers were added under continuous stirring in an oxygraphic cell containing R-buffer (2.77 mM CaK 2 EGTA, 7.23 mM K 2 EGTA (100 nM free Ca 2+ ), 1.38 mM MgCl 2 (1 mM free Mg 2+ ), 20 mM taurine, 0.5 mM dithiotreitol, 90 mM potassium-methane sulfonate, and 20 mM imidazole, 10 mM sodium-methane sulfonate and 2 mg/mL bovine serumalbumine, pH 7.1) and malate/glutamate (4 mM/10 mM). O 2 consumption was measured in the absence (state 2) and the presence of 2 mM adenosine diphosphate (ADP) (state 3 with complex I-linked substrates). Then, 2 mM amytal and 10 mM succinate were added (state 3 with complex II-linked substrates). O 2 consumption rates are given in micromoles of O 2 per minute per mg of proteins. The acceptor control ratio (ACR), an index of oxidation-phosphorylation coupling, was also calculated as the ratio of state 3 to state 2.
For cardiac mitochondrial ATP production, freshly isolated subsarcolemmal (SSM) and interfibrillar (IFM) mitochondria were incubated at 25 • C in R-tampon with 10 mM glutamate and 4 mM malate for 10 min under continuous stirring to eliminate sample residual ADP. Then, 2 mM ADP was added and samples were collected as previously described [13]. The rate of ATP production was evaluated by a bioluminescence assay kit (Roche Diagnostics, France) according to the manufacturer's protocol. Values were normalized to the protein concentration of each sample.

Statistical Analysis
All values are expressed as means ± standard error of mean (SEM). Student's t-test was performed to compare means between control and NO 2 exposed groups. Statistical analyses were performed using GraphPad Prism (version 7.04, GraphPad software, USA). Differences were considered statistically significant when p < 0.05.

Results
3.1. Acute NO 2 Exposure Induced a Rapid but Reversible Cardiac Response LV function was evaluated by echocardiography after acute exposure to NO 2 (Figure 2A-D). Relative to control group, single exposure to NO 2 induced a significant increase in LV diastolic and systolic diameters (10 and 38%, respectively), and a decrease in fractional shortening (−23%), after 1 h of recovery period. This effect appeared transient since a 24 h recovery period post-acute exposure erased the increase in LV diastolic and systolic diameters and the decrease in fractional shortening were no longer observed. Cardiac output ( Figure 2D) remained unchanged after NO 2 exposure, whatever the recovery period.

Statistical Analysis
All values are expressed as means ± standard error of mean (SEM). Student's t-test was performed to compare means between control and NO2 exposed groups. Statistical analyses were performed using GraphPad Prism (version 7.04, GraphPad software, USA). Differences were considered statistically significant when p < 0.05.

Acute NO2 Exposure Induced a Rapid but Reversible Cardiac Response
LV function was evaluated by echocardiography after acute exposure to NO2 (Figure 2A-D). Relative to control group, single exposure to NO2 induced a significant increase in LV diastolic and systolic diameters (10 and 38%, respectively), and a decrease in fractional shortening (−23%), after 1 h of recovery period. This effect appeared transient since a 24 h recovery period post-acute exposure erased the increase in LV diastolic and systolic diameters and the decrease in fractional shortening were no longer observed. Cardiac output ( Figure 2D) remained unchanged after NO2 exposure, whatever the recovery period.  To gain further insight into the putative impact of NO 2 exposures on vascular function, we next investigated coronary endothelial function. Coronary arteries were isolated from the hearts of control and NO 2 -exposed rats for vascular reactivity assessment, as a measure of endothelium-dependent coronary vasodilatation after serotonin pre-constriction and the addition of acetylcholine. Acute NO 2 exposure did not impair endothelial function in coronary arteries, as shown by comparable endothelial relaxation in both control and acute NO 2 exposure groups (Figure 3). To gain further insight into the putative impact of NO2 exposures on vascular function, we next investigated coronary endothelial function. Coronary arteries were isolated from the hearts of control and NO2-exposed rats for vascular reactivity assessment, as a measure of endothelium-dependent coronary vasodilatation after serotonin pre-constriction and the addition of acetylcholine. Acute NO2 exposure did not impair endothelial function in coronary arteries, as shown by comparable endothelial relaxation in both control and acute NO2 exposure groups (Figure 3).

Acute NO2 Exposure Induced Rapid but Transitory Cardiac Mitochondrial ROS Production without Mitochondrial Dysfunction
To investigate mitochondrial function after NO2 exposure, the OXPHOS capacity was investigated in situ from permeabilized cardiac fibers in order to maintain the mitochondria in their normal intracellular position and their interactions with other organelles. Acute NO2 exposure did not affect the mitochondrial function, since ADP-independent respiration with glutamate and malate (state 2) was similar between the groups, as well as ADP-dependent respiration with glutamate and malate (state 3 complex I) or succinate (state 3 complex II) as substrates ( Figure 4A). In these conditions, oxidation-phosphorylation coupling, determined by the ratio of respiration rate before and after the addition of ADP (acceptor control ratio ACR, state 3 (complex I)/state 2), remained unchanged ( Figure 4B).
To further explore the effect of acute NO2 exposure on mitochondrial function, we next measured both ATP production rates and superoxide production in cardiac mitochondrial subpopulations, subsarcolemmal (SSM) and interfibrillar (IFM) mitochondria. Acute NO2 exposure did not alter ATP production rates, confirming the absence of a mitochondrial dysfunction ( Figure  4C), although a temporary increase in mitochondrial ROS production was observed specifically in IFM. Indeed, ROS levels were similar between the groups, 1-day post-exposure ( Figure 4D).

Acute NO 2 Exposure Induced Rapid but Transitory Cardiac Mitochondrial ROS Production without Mitochondrial Dysfunction
To investigate mitochondrial function after NO 2 exposure, the OXPHOS capacity was investigated in situ from permeabilized cardiac fibers in order to maintain the mitochondria in their normal intracellular position and their interactions with other organelles. Acute NO 2 exposure did not affect the mitochondrial function, since ADP-independent respiration with glutamate and malate (state 2) was similar between the groups, as well as ADP-dependent respiration with glutamate and malate (state 3 complex I) or succinate (state 3 complex II) as substrates ( Figure 4A). In these conditions, oxidation-phosphorylation coupling, determined by the ratio of respiration rate before and after the addition of ADP (acceptor control ratio ACR, state 3 (complex I)/state 2), remained unchanged ( Figure 4B).
To further explore the effect of acute NO 2 exposure on mitochondrial function, we next measured both ATP production rates and superoxide production in cardiac mitochondrial subpopulations, subsarcolemmal (SSM) and interfibrillar (IFM) mitochondria. Acute NO 2 exposure did not alter ATP production rates, confirming the absence of a mitochondrial dysfunction (Figure 4C), although a temporary increase in mitochondrial ROS production was observed specifically in IFM. Indeed, ROS levels were similar between the groups, 1-day post-exposure ( Figure 4D).

Repeated NO2 Exposures Impaired Cardiovascular Responses
We next wanted to determine whether the duration of NO2 exposure has differential effects on cardiovascular function. Three weeks of NO2-exposures caused a sustained cardiac effect, since 24 h after the 3-week exposure, LV diastolic and systolic diameters (−12 and 34%, respectively) as well as LV fractional shortening and cardiac output (−20% and −17%, respectively) were severely altered compared to time-matched controls ( Figure 5).

Repeated NO 2 Exposures Impaired Cardiovascular Responses
We next wanted to determine whether the duration of NO 2 exposure has differential effects on cardiovascular function. Three weeks of NO 2 -exposures caused a sustained cardiac effect, since 24 h after the 3-week exposure, LV diastolic and systolic diameters (−12 and 34%, respectively) as well as LV fractional shortening and cardiac output (−20% and −17%, respectively) were severely altered compared to time-matched controls ( Figure 5).
In parallel with this cardiac dysfunction, repeated NO 2 exposures impaired endothelium-dependent relaxation, illustrated by the decrease of coronary relaxation in the repeated NO 2 exposure group ( Figure 6A). No differences were observed between groups upon sodium nitroprusside (SNP)-induced relaxation ( Figure 6B). In parallel with this cardiac dysfunction, repeated NO2 exposures impaired endotheliumdependent relaxation, illustrated by the decrease of coronary relaxation in the repeated NO2 exposure group ( Figure 6A). No differences were observed between groups upon sodium nitroprusside (SNP)induced relaxation ( Figure 6B). Incubation of coronary arteries with the NOS inhibitor L-NNA markedly reduced the relaxing response in both groups ( Figure 6C). Incubation of coronary arteries with superoxide dismutase did not modify acetylcholine responses in the control group but improved the impaired relaxation observed after repeated NO2 exposure ( Figure 6D). Incubation of coronary arteries with the NOS inhibitor L-NNA markedly reduced the relaxing response in both groups ( Figure 6C). Incubation of coronary arteries with superoxide dismutase did not modify acetylcholine responses in the control group but improved the impaired relaxation observed after repeated NO 2 exposure ( Figure 6D).

Repeated NO 2 Exposures Impaired Mitochondrial Function and Mitochondrial ROS Production
In order to determine whether repeated NO 2 exposure affected mitochondrial function, we performed exposures during three weeks and the assessments were performed at 1-day post-exposure.  Respiration with either complex I or complex II substrates was reduced by 21 and 23%, respectively ( Figure 7A) after repeated NO 2 exposure. In line with these results of state 3 respiration, ACR was lower in NO 2 -compared to air-exposed rats ( Figure 7B). The ATP production rate decreases observed in both SSM and IFM fractions confirmed this mitochondrial dysfunction ( Figure 7C). An increase in mitochondrial superoxide production was observed specifically in IFM ( Figure 7D) and was correlated with the decrease in cardiac output ( Figure 7E).

Discussion
The results of this study show that acute NO 2 exposure induced an early, but transitory mitochondrial superoxide production associated with reversible impairment of cardiac function, whereas repeated exposures induced mitochondrial and cardiac dysfunctions, which persisted for 24 h after the last exposure. Moreover, repeated NO 2 exposures impaired ACh-mediated dilatation in coronary arteries, an effect that was due to a decrease in NO bioavailability caused, at least partially, by increased superoxide production. Taken together, these data provide the first evidence that NO 2 exposures impaired cardiac mitochondrial function, which, in conjunction with coronary endothelial dysfunction, contributes to a sustainable cardiac dysfunction.
Air pollution has long been recognized as a major CV risk, due particularly to fine particulate matter (PM 2.5 ) derived from combustion sources, and road traffic and diesel exhaust (DE). However, the potential effects of a co-pollutant gases, such as NO 2 , has been neglected. In a previous study, we demonstrated that removing particles from the DE aerosol did not protect against DE-induced cardiac and mitochondrial dysfunctions, revealing an important implication of the gaseous phase in this response [13]. In this present study, acute NO 2 exposure at a concentration close to that measured in DE induced a cardiac impairment characterized by an LV dilatation and associated with reduced LV fractional shortening, reflecting a possible loss of contractility. This LV dysfunction appeared only moderate since cardiac output was not modified. Moreover, this effect was reversible since, after 1 day post-exposure, echocardiographic parameters were comparable to those measured in control group. However, cardiac dysfunction persisted and worsened after repeated exposure to NO 2 over three weeks. These evaluations were made at day-1 post-exposure, excluding the deleterious direct effect of NO 2 as observed 1 h after acute NO 2 exposure. These results are consistent with a cohort study that has shown an association between past exposure to NO 2 and PM 2.5 and cardiac ventricular dilatation [25], a marker of adverse remodeling that often precedes heart failure development, stressing the irreversible effects of these exposures on the heart.
To identify the underlying cellular mechanisms involved in the cardiac dysfunction following NO 2 exposure, we first focused on cardiac mitochondrial ROS and ATP production. Two spatially distinct mitochondria subpopulations have been observed in myocytes and may be associated with a specific response to pathological stimuli, indicating the role of IFM in contractile function [26,27]. The present study revealed a primary but reversible interfibrillar mitochondria-specific increase in ROS after acute exposure. This increase was not accompanied by a mitochondrial alteration, as evidenced by the maintenance of ATP synthesis capacity. The increase in mitochondrial ROS can reflect an adaptive event involved in cardio-protection, as previously demonstrated [28,29]. For example, Antonucci et al. [29] investigated the effects of mitochondrial ROS production using a mitochondria-targeted redox cycler MitoParaquat (MitoPQ) and showed that low levels of ROS are cardioprotective, while higher doses of MitoPQ resulted in a progressive alteration of mitochondrial function in vitro. The magnitude but also the repetition of the mitochondrial ROS production may result in the alteration of mitochondrial function and in worsening of cardiac function, as observed after repeated NO 2 exposures. The strong correlation between mitochondrial ROS production and cardiac output associated with the appearance of mitochondrial dysfunction support this hypothesis. The fact that mitochondrial ROS production precedes the mitochondrial dysfunction suggest that this production is an early trigger event in the onset of mitochondrial defect. Indeed, we showed that respiration rate was affected in the hearts of rats exposed to repeated NO 2 exposures. This is consistent with our previous study showing that repeated DE exposures induced an impairment of mitochondrial function associated with deficient cardiac contractility [13]. With regard to other combustion-related gas, CO [30] and SO 2 [31] exposures also induced cardiac mitochondrial effects. Overall, these results suggest that mitochondria impairment contributes to the CV events linked to air pollution.
Then, we focused on coronary endothelial dysfunction. Indeed, despite the number of studies showing that air pollution, and more specifically DE particles, causes endothelial dysfunction [32][33][34], the contribution of the gaseous phase in these effects is underestimated and the effect of NO 2 itself on endothelial function is unknown. After acute exposure, we did not observe any effect on endothelial function. This is in line with a previous study showing no impact on brachial artery reactivity after 1 h exposure of humans to 4 ppm NO 2 [8]. Nevertheless, repeated NO 2 exposures impaired relaxation to acetylcholine in coronary arteries from healthy rats. Endothelial NO participates in the control of vascular tone and changes observed in the reactivity of coronary arteries after NO 2 exposures might be explained by a decrease in NO availability. As repeated NO 2 did not modify relaxation induced by sodium nitroprusside, a NO donor, a direct effect of NO 2 on vascular smooth muscle cells is unlikely. These results demonstrate that NO 2 exposures impaired coronary endothelial cell function, an effect that was due to a reduction of endothelial NO availability. One possible mechanism of reduced NO availability might be explained by an increased superoxide production in arteries after NO 2 exposures, since superoxide scavenging by SOD restored attenuated ACh-induced relaxation of coronary arteries. The reduction in NO-bioavailability and the resulting altered coronary function probably contributes to a decrease in myocardial perfusion, which is likely to contribute to hypoxia-induced ROS production and, as a probable consequence, causes the vicious circle of ROS-induced ROS release [35,36]. Although we did not evaluate the pro-inflammatory markers in the present study, we cannot exclude a contribution of circulating factors in these effects. A previous study suggested vascular toxicity mediated by pro-inflammatory circulating factors following exposure to NO 2 [12]. These authors showed that plasma from humans exposed to 0.5 ppm NO 2 was able to activate expression of cell adhesion molecules by coronary endothelial cells in culture [12]. Further studies are needed to determine the precise underlying mechanism.
In conclusion, we demonstrate for the first time that repeated NO 2 exposures altered cardiac mitochondrial and cardiac function, as well as coronary vascular reactivity, because of endothelial dysfunction. Moreover, these results demonstrated that acute exposure to NO 2 induced a mitochondrial ROS production, which could represent an early trigger event in the onset of cardiovascular defects after chronic exposures. Funding: This research was funded by ADEME (French Environment and Energy Management Agency; Cardiox project, grant number 2013-1-237), and co-supported by the Rouen Normandie University and the European Union (Europe gets involved in Normandy with European Regional Development Fund; ERDF). Ahmed Karoui was a receipt of a PhD grant from the ADEME and the Regional Council of Haute-Normandie, Clément Crochemore was a receipt of a PhD grant from the Regional Council of Haute-Normandie and Najah Harouki was a receipt of a PhD grant from the Rouen Normandie University. The authors thank F. Dionnet, V. Keravec and E Estace from CERTAM (Saint-Etienne du Rouvray, France) for in vivo facilities.